Genetics of amyotrophic lateral sclerosis

Key points to know

Although the pathophysiology of multiple side amyotro-phic (ALS) remains unknown to this day, there is a consensus that the involvement of genetic factors is a key element in the death of motor neurons. Since 1993 and the discovery of mutations in the SOD1 gene , the number of genes linked to or promoting the occurrence of ALS has grown steadily . 

Among them, only 4 ( SOD1, TARDBP, FUS, C9ORF72 ) are unanimously considered as major pathogenic genetic factors of ALS, responsible for just over two thirds of familial forms and may be more than 5 % of forms considered sporadic. A major parameter in the gene relationship -ALS appears to be more and more often the geographical origin of the population studied, this seems particularly marked for the C9ORF72 gene . 

For genetic susceptibility factors, the increased risk of developing ALS is clearly established for the SMN1 and ATXN2 genes , and is accepted for certain haplotypes of the VEGF gene . 

Modulating effects, linked to populations, also seem to exist, as underlined by the different ApoE – SLA relationships between European and North American studies. In possibly hereditary ALS, the European Neurology Federation has published recommendations that limit the search for responsible genes and hereditary forms to very specific situations. 

Advances in the field of genetics also raise questions about the relevance of the exploration of asymptomatic relatives of ALS patients, which in particular justifies that the mutation present in the patient is well identified as a pathogenic mutation.

B hile amyotrophic lateral sclerosis (ALS) is currently the third neurodegenerative disorder, in frequency, in adults, the pathophysiology of this devastating disease is still unknown. It is however accepted by all that its pathogenesis is multifactorial. Among all the avenues explored, the involvement of genetic factors has been the subject of a great deal of work, in particular over the past twenty years since the description of mutations in the superoxide dismutase 1 gene.

From ALS to genetics: 100 years of history . . .

Although the existence of genetic factors in ALS is now fully recognized, the relationships between genetics – inheritance – ALS have long been controversial. Indeed, it took almost 80 years after the original description of J.-M. Charcot for the concept of familial forms of ALS to be recognized, thanks to the work of Mulder and Kurland in 1955. 

This recognition then led to an attempt to clarify the phenotype of familial forms of ALS, which was established 30 years later. If it is admitted that it is not possible to distinguish individually a sporadic form from a family form, the latter have certain characteristics which make it possible to differentiate the patient populations: family forms have on average a early 10 years earlier, sensory disturbances are not unusual and the rate of evolution follows a bimodal curve . 

Since 1993 and the identification of mutations in the SOD1 gene , nearly twenty genes have been associated with this condition . The objective of this review is to take stock not only of pathogenic genes but also of those of susceptibility to ALS. We will emphasize the factors for which the link or association with ALS is clearly established. Pathogenic genes correspond to genes for which causal mutations are reported and accepted. Several elements must first be checked to retain the impli-cation of a gene in ALS: the mutation must co- segregate with the disease in families and this mutation must not be found in control subjects. 

Genes whose pathogenesis is established gene

for superoxide dismutase 1 gene SOD1 was the first identified genetic factor in ALS, it is located on 21q22.1 locus . To date more than 160 mutations have been described within the 5 exons, they are mostly dominant . Only the D90A and D96N mutations are of recessive transmission . Mutations in the SOD1 gene are mainly missense mutations. 

The mutated SOD1 protein would exert its pathogenic action by gaining function: thus, by allowing to catalyze the formation of the peroxynitrite ion , it will cause the nitration of proteins (including neurofilaments), and therefore the alteration of the cytoskeleton and flow axonal.

Clinically, the patient profile has certain particularities: an earlier onset age than that of sporadic forms of approximately 10 years, a preferential onset in the lower limbs (70 to 80% of cases) and the absence of cognitive impairment. It is important to underline a great heterogeneity in the duration of evolution, which goes from less than a year for the A4V mutation, to more than 20 years for the D90. 

The phenotypic study of patients also showed great clinical heterogeneity of ALS tables mutation to another, but also in the presence of the same mutation even within a family; the presence of sensitive symptoms, for example, is not exceptional, the absence of pyramidal signs also characterizes certain mutations or families . Currently, the D90A mutation is the most clinically studied SOD1 mutation and is the most common in the world. 

Movement disorders always start in the lower limbs, after a phase ” pre paretic ” characterized by cramping and myalgias, and painful paresthesia of the lower limbs. The hole-ble sphincter have also been described, t out as with the G41S mutation . It is possible to draw a map of SOD1 mutations which seem for the most part specific to a geographic region: thus, the A4V mutation predominates in subjects of Italian origin, the I112M mutation is more frequent in Hispanics, the R115G mutation more often observed in Germanic people and the D90A mutation very prevalent in Sweden (1 in 17 individuals).


The TARDBP ( Tar DNA-binding protein ) gene , located on locus 1p36.22, was linked to ALS in 2009 . It codes for a 43 kD protein called TDP-43. It is involved in several metabolic processes of DNA and RNA including DNA repair, regulation of transcription and splicing of RNA. 

The demonstration of mutations on this gene was supported by the existence of inclusions marked by anti-TDP-43 antibodies in the cytoplasm of neonates of ALS patients. Current LEMENT, the TARDBP mutations are present in 5% of the familial cases and less than 1% of the sporadic forms. 

Although recently discovered , nearly 50 different mutations have been described in this gene. Unlike the SOD1 mutations , subjects carrying a mutation in the TARDBP gene have an ALS that begins more often in the upper limbs (66%), except in Asia where the initial picture is frequently bulbar. 

The age of onset is also 10 years earlier than that of sporadic forms and the median survival is double that of sporadic forms . Although not usually associated with cognitive impairment, TARDBP mutations have also been described in frontotemporal dementia (DFT) or in forms of ALS with cognitive impairment .

Gene fused in sarcoma (FUS / TLS)   

This gene is located on locus 16p11.2 . FUS has many functional similarities with the TARDBP gene including a strategic region rich in glycine and two RRM regions (RNA recognition motif ) which bind DNA and RNA, playing a role in DNA repair, as well as in regulation of RNA transcription and splicing. 

Mutations in the FUS gene are linked to less than 5% of familial forms and seem to be anecdotal in sporadic forms . These mutations are all dominant except one reported in a Cape Verde family of recessive transmission. FUS mutations are linked only to ALS and have never been reported in DFT . An analysis of family trees points to a high penetrance of the mutation illustrated by families with many cases of ALS. 

Currently, nearly 60 mutations have been described, the most frequent affecting exon 14 and more specifically codon 521 (arginine). The clinical picture characteristic of these mutations seems to be that of an ALS starting in 40% of the cases before the age of 40 by a motor deficit of the extensors of the cervical spine, predominant on the peripheral motor neuron ( NMp ), without bulbar involvement and cognitive, and evolving rapidly in less than 2 years . 

Juvenile mutated FUS forms with an age of onset of symptoms around the age of 11 have recently been described, but these exceptional situations remain controversial .

Gene C9ORF72 

This gene, which is one of the most recently described, is currently the one whose mutations account for the largest number of cases of familial ALS. It is implicated in 40% of familial forms and in 5 to 20% of falsely sporadic forms depending on the population; A North-South gradient is emerging with a preponderance of mutated cases in the Nordic populations due to a North European founding effect. 

Certain populations seem relatively untouched by this gene: this is the case of India and Australia, probably due to the low mix of these populations with individuals from northern Europe . Recent results suggest that this founding effect dates back 1,500 years, more than 100 generations ago . In France, the studies conclude that frequency of 46% of ALS fami lia them and 8% of sporadic ALS . 

The phenotype of the patients carrying this mutation seems to be able to be specified: the age of onset is earlier by 5 years on average in this population, but the penetrance is not complete until after 80 years, explaining the cases of pace sporadic , ascendants who may have died from other causes before the onset of ALS. The precociousness of age of onset increases over generations with an age of onset of ALS on average 7 years earlier in the following generations. The predominance of spinal or bulbar form remains debated according to the populations studied. Finally, the progression of the handicap is faster and the duration of evolution shorter than in the classic forms of ALS . 

The authors point out the very frequent presence of cognitive or behavioral disorders in cases linked to a C9ORF72 mutation , which also distinguishes these forms from those linked to a TARDBP or SOD1 mutation in which cognitive disorders are observed in approximately 30% and 2 % of patients, respectively . Cerebral imagery would have certain specificities in this population: in fact, atrophy predominates on the frontal orbito regions , while in the DFT linked to a MAPT or GRN mutation it predominates respectively on the anterior temporal regions and in temporo-parietal . However, it should be emphasized that C9ORF72 mutations are also the most frequent in familial DFTs. 

They thus appear as the missing link in the search for a continuum between ALS and dementia, which has been the subject of debate for the past 30 years. The exact role of the mutated protein in the mechanisms of motor neuron degeneration remains unknown, oscillating between haplo-insufficiency or gain in function. 

For the first time in the genetics of ALS, an algorithm has been proposed in order to better target the search for C9ORD72 mutations ; the existence of a family history of DFT and the existence of cognitive disorders of the behavioral line in a patient are predictors of a mutation in this gene . 

The anomalies described with this gene are different from those reported with the SOD1 , TARDBP and FUS genes : they are not mutations in the coding sequence but an abnormal repetition of a 6 nucleotide sequence (a GGGGCC hexaplet) located in an intronic sequence , that is to say non-coding gene. 

In recent months, several studies have suggested that the clinical phenotype, in the presence of the C9ORF72 mutation , is more complex than expected. Indeed, cases of Parkinson’s disease, corticobasal degeneration , cerebellar syndromes have been described. 

The situation therefore still needs to be deepened in order to really know the spectrum of neurological damage covered by these mutations. In total, more than two thirds of familial forms of ALS are linked to a mutation in one of the 4 genes described above .

Potentially pathogenic genes

Valosin-containing protein ( VCP ) gene

The VCP gene is located on locus 9p13.3. The VCP mutations were initially reported in patients with a picture characterized by the clinical triad ” inclusion myositis – Paget’s disease – frontotemporal dementia ” , known by the acronym IBMPFD, in thirty families worldwide . 

These mutations were then described in classic forms of ALS, without an increase in alkaline phosphatases and without any sign of Paget’s disease . To date, the VCP – SLA relationship remains to be clarified.

Gene for optineurin ( OPTN ) 

The OPTN gene is located on the 10p13 locus. The transmission of this mutation is either dominant or recessive. The implication of the OPTN gene remains controversial, as several studies have not confirmed the original results . 

One explanation for this discrepancy could once again be geographic, this mutation appearing to be linked to ALS in Asian and non-Caucasian populations. However, we cannot exclude that the princeps publication is erroneous, since the families concerned by the work published in Nature , had glaucoma and that the OPN gene is linked to this pathology.

Ubiquilin 2 gene ( UBQLN2 ) 

The UBQLN2 gene is carried by the X chromosome. The ALS linked to the UBQLN2 mutations are X-linked dominant transmission. The relationship between this gene and ALS relates in particular to the co- segregation of the disease with the mutation and the presence of UBQLN2 + inclusions in the spinal cord of patients. 

The phenotype is more severe in men with an earlier onset age of 15 years (33.9 versus 47.3 years), the duration of evolution is similar between the sexes . However, to date, the initial results of de Siddique et al. have not been replicated.

Gene vesicle associated membrane protein ( VAPB )

This gene is carried by locus 20q13.32. The mutations relate almost exclusively to patients of Brazilian origin. Three phenotypes are linked to this gene: a classic form of ALS, an atypical form characterized by the association with the attack of the motor neuron of a postural tremor and an unusual slowness of evolution and, finally, a form d spinal muscular atrophy in adults. The most common mutation is P56S . 

To date, these mutations remain rare in ALS and the causal link remains discussed; the search for a VAPB mutation in ALS does not seem indicated except in very specific cases .

Susceptibility genes 

Susceptibility genes are genes more frequently found in ALS than in the general population. They increase the risk of having the disease. The first genetic susceptibility factor described was the ApoE gene . 

Several studies have studied the frequency of the various alleles in ALS, making it possible to objectify an increased risk of bulbar ALS in patients carrying the ApoE4 allele, while the ApoE2 allele seems to predominate in early spinal forms . Once again, a population effect seemed to modulate these results, the association being found only in European studies and not in North American ones.

Motoneuron survival genes 

The motor neuron survival gene ( SMN ) is carried by the 5q13 locus. The main characteristic of this locus is a mirror duplication of a region of 500 kb which explains the presence on each of the chromosomes of 2 SMN copies . The SMN1 telomeric copy is that responsible for infantile spinal muscular atrophy (ASI) in the event of homozygous deletion of exons 7 and / or 8.

The centromeric copy has a lesser role in modulating the phenotype: the least Severe ASIs are linked to a higher number of SMN2 copies . The SMN protein is involved in many cellular activities including that of RNA metabolism. Interestingly, its role seems close to that of TDP and FUS described above. Given the strong links between SMN and ASI, this gene appeared very quickly as a major genetic candidate in ALS. 

Several studies have shown that an abnormal number of SMN1 copies (1 or 3 copies) significantly increases the risk of developing ALS . Although there is not, unlike what is reported in the ASI, a modulating effect of the number of copies of SMN2 on the duration of evolution, there is however a longer evolution in patients Swedish carriers of a homozygous SMN2 deletion , compared to that of French patients carrying such a deletion, once again showing an influence of the geographical origin, and therefore of the genetic capital of an individual, in the modulation of the action of other genes .

Paraoxonase ( PON ) gene  

PON is an enzyme that protects low density lipoproteins against oxidation and plays a detoxifying role against various products such as drugs, organophosphates and neurotoxins. There are 3 genes encoding a PON on locus 7q21, the detoxifying activity being carried essentially by PON1; if this association remained conjectural until recently despite 6 studies in favor of a link (not confirmed by meta-analyzes), a sequencing study of the 3 PON genes in 260 familial ALS and 188 sporadic ALS allowed ” identify 8 mutations among which 5 affected PON1 in the region coding for enzymatic activity . 

Although the effect of these mutations on protein function remains to be determined (gain or loss of function), their identification supports the hypothesis of lipid oxidation as one of the physiopathological mechanisms involved in ALS.

Gene of the vasculo endothelial growth factor ( VEGF ) 

Work on the VEGF gene has had a significant impact on the pathophysiological hypotheses of ALS by supporting the relationships between hypoxia and death of the motor neuron. Indeed, VEGF plays a major role in the tissue response to hypoxia: the deletion in the VEGF promoter region of the HRE region ( hypoxia -responsive element ) in mice leads to degeneration of motor neurons. 

Several studies have sought to support VEGF as a candidate gene for ALS. Although a study in a large population has shown a link between certain VEGF haplotypes and ALS, no study to date has identified VEGF mutations in ALS .

Angiogenin gene 

In connection with the theory of hypoxia and VEGF , the angiogenin ( ANG ) gene has also been the subject of various works: several mutations of the ANG gene have been described in familial and sporadic forms of ALS of different er geographic origins . These studies therefore strengthen the relationships between hypoxia and death of the motor neuron in ALS.

Sequestosome 1 gene ( SQSTM1 ) 

This gene is carried, like the SMN gene , by the 5q locus. SQSTM1 encodes the p62 protein involved in the processes of autophagy and in response to oxidative stress. Mutations in this gene have been reported in ALS and also in DFT . Insofar as SQSTM1 codes for the protein p62 which is involved in the response to oxidative stress, these results support the hypothesis of the role of oxidative stress in ALS.

Genes Modifying the ALS Phenotype

Gene of ataxin 2 

The ataxin 2 gene ( ATXN 2) is carried by the locus 12p24.12. It was initially linked to a completely different picture of ALS, type 2 spinocerebellar atrophy (SCA2). This condition is linked to an abnormal repetition of a CAG triplet exceeding 34 copies ( n _ 23). 

In ALS, the number of repetitions is abnormally high but in lower values ​​than those linked to SCA2 (number of repetitions between 24 and 34), and there would be a correlation between the number of repetitions and the severity of ALS. The pathogenic role of ATXN2 in ALS is explained by an effect promoting the aggregation of the protein TDP -43 in motor neurons in cell and animal models .

Locus 1p34  

Last year, a new region of interest was identified in ALS: locus 1p34. A meta-analysis carried out on behalf of the ALSGEN consortium has shown an association with this region: in fact, patients with the allele start their disease 2 years earlier than those with the AA haplotype .

EPHA4 gene 

The EPHA4 gene is located on the X chromosome. The EPHA4 protein, an ephrine , inhibits the process of re-innervation of motor neurons. The gene deletion EphA4 in Mode-the mouse SOD1 is associated with a progressive profile slowed animals. This effect is more dose-dependent since the strains carrying a homozygous deletion of EPHA4 have a longer evolution than that of the strains having a heterozygous deletion (146 days vs 135 days, p = 0.0063). 

In ALS, mutations in the EPHA4 gene have been described. These cases of ALS were all characterized by a particularly slow evolutionary profile, beyond 90 months of survival, joining the descriptions made in animals and the supposed effect of ephrine a4 . However, to date, no large study has replicated these results. 

Genome wide screening , a technology for the future ?

A more recent approach to identifying genetic risk factors has been that of a comprehensive genome study. This approach, which allows the exploration of hundreds of thousands of genetic markers on a large number of samples, requires international recruitment; this leads to an analysis bias linked to the heterogeneity of the geographic origin of the sampled patients. This could explain the poor reproducibility of the results .

Can we diagnose ALS at a pre-symptomatic stage by genomic exploration?

One of the recurring fears of patients with ALS is the risk of transmission of this pathology to their descendants. This legitimate concern requires precise information. Indeed, with regard to pre-symptomatic diagnoses, research must be limited to direct relatives of patients carrying a mutation already known as a pathogen. 

Furthermore, it is important to remember about primarily diag-nosis pre-symptomatic that the identification of a mutation does not guarantee necessarily develop ALS, firstly, and, secondly, it will it is not possible to predict the possible age of onset or even the clinical severity . These elements are particularly important for C9ORF72 mutations because their complete penetration is very late (more than 80 years).

Can we define a genetic analysis strategy in ALS?

While it is clear that there are currently no rules for looking for mutations in an ALS patient, two parameters seem to dictate our attitude: _ what situation justifies genetic research? _ which genes should we study in priority? The European Federation of Neurological Societies (EFNS) has clarified the situations in which a genetic study was justified in ALS: this research is indicated in familial forms but also in sporadic forms suggestive of a recessive D90A mutation. 

The classic forms of ALS do not warrant genetic study. It is also desirable to prioritize genetic research based on the frequency of mutations in ALS and especially based on the phenotype. Since the 4 major genes, C9ORF72, SOD1, TDP and FUS , represent more than two thirds of cases of hereditary ALS, their study should be carried out as a priority. 

However, given the frequency of the two main mutations, the SOD1 and C9ORF72 genes must be studied as a first intention, especially since there is an onset in the lower limbs ( SOD1 ) or psycho-behavioral disorders in the patient or among its relatives ( C9ORF72 ). In case of negativity of this research or in the event of a start in the upper limb, this will rather encourage the search for a mutation of the TARDBP gene while a story starting before the age of 40, rapidly evolving, in a family with high penetration will rather lead to a FUS mutation . 

Molecular biological analysis of other genes should be done on a case-by-case basis, given the low frequency of mutations reported in ALS. If a mutation was thus discovered, it will then be necessary to ensure its pathogenic character by its co- segregation with ALS within the family (the fact that only ALS carries the mutation and that non-sick subjects do not carry it ) and possibly by verifying its absence from a representative control population. The study of ALS susceptibility genes, for its part, is not justified for clinical purposes to date, its interest remains purely scientific.

What we must keep in minde

The genetics of ALS have seen great advances in the past twenty years: now more than two-thirds of hereditary familial forms have an identified genetic cause. 

These advances do not mean that the gene – ALS relationships are fully established; an environmental factor or other gene linked to the origin geogra-phic certainly modulates the pathogenic effect with specific-ficités mutations ( eg . SOD1 , TARDBP ) or mutated genes ( UBQLN2, OPTN ) in different populations. 

This surely requires a better definition of the genotype – phenotype relationships and the planning of analyzes in homogeneous groups of patients; the results obtained with the SOD1 gene and the description of phenotypes specific to certain mutations illustrate the relevance of such an approach.